Additional file 9: of Sequence-based prediction of protein protein interaction using a deep-learning algorithm

Detailed analysis of the prediction accuracies of the test sets. Table S6. Percent of proteins in the test sets having ≥30% sequence identity to those in pre-training/whole benchmark dataset and the prediction accuracy. Figure S5. Relationship between the prediction accuracy and the percent of proteins in a test set with ≥30% sequence identity to those in the training set. (DOCX 105 kb